Skip to content

Architecture Modification / Research · Jupyter Notebook

VISTA

VISTA: Velocity-Informed Staticness Temporal Attention A Physics-Aware Motion Prior for Stable Object Detection in Automotive Perception

What it is

VISTA is a video object detection framework that augments standard detection architectures (DETR, YOLO, SSD) with optical flow-based motion features and a staticness branch to improve detection of moving and static objects. It provides training scripts for three variants and includes inference code.

Scored from the code at its last scan, 8 Oct 2026. How merit is scored

Where it counts

  • Machine Learning90%
  • Computer Vision90%
  • Data Science20%

How much of the project is about each field, as scoring counts it. It ranks in the field it is most about.

Around it on the Machine Learning board

Whole board
  1. 33AgroADASML Application · Python52.6/100ImplementationSuceveanu Mircea Dan Sandu
  2. 34ml-from-scratchEducational Implementation · Jupyter Notebook51.9/100ImplementationPop Alexandru
  3. 35VISTAArchitecture Modification / Research · Jupyter Notebook50.4/100ImplementationSuceveanu Mircea Dan Sandu
  4. 36chatbot-pipelineML Application · Python50.2/100ImplementationPop Alexandru
  5. 37socrateLibrary / Framework · Jupyter Notebook49.2/100Implementationvlad andrei

More from Suceveanu Mircea Dan Sandu